菜单
Use when implementing reactive fault-tolerance patterns in Spring Boot 4 with native Spring Framework 7 resilience features or Resilience4j, including retries, concurrency limiting, circuit breakers, rate limiting, timeouts, and fallbacks for downstream calls.
默认使用会先检查来源的 Prompt；你也可以切换为直接命令，或下载本地副本。
决定是否安装前，请先阅读 SKILL.md，以及 SkillsMP 当前展示的配套文件。
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
直接命令不会经过审查 Prompt；运行前请先检查来源。
npx skills add https://github.com/dallay/profiletailors.com --skill spring-boot-resilience命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地？可下载 SkillsMP 当前能够提供的文件。
基于 SOC 职业分类
正在显示 SKILL.md
| name | spring-boot-resilience |
| description | Use when implementing reactive fault-tolerance patterns in Spring Boot 4 with native Spring Framework 7 resilience features or Resilience4j, including retries, concurrency limiting, circuit breakers, rate limiting, timeouts, and fallbacks for downstream calls. |
| allowed-tools | Read, Write, Edit, Bash |
Resilience patterns for Spring Boot 4 + Spring Framework 7 + WebFlux + Kotlin coroutines.
This skill covers the official resilience model now available in Spring Framework 7, plus the cases where Resilience4j still makes sense for more advanced control.
The official documentation confirms that Spring Framework 7 includes native resilience support:
@Retryable is built in@ConcurrencyLimit is built in@EnableResilientMethods@Retryable applies to reactive return types and decorates the Reactor pipelineFor reactive HTTP clients in Spring Boot 4, the official client baseline is:
WebClientWebClient.BuilderUse companion skills instead when the main concern is:
spring-boot-api-standardsspring-bootspring-boot-securityspring-boot-testing-webflux /
spring-boot-testing-integrationsWebClient, not RestTemplate.Use @EnableResilientMethods in configuration to activate Spring's native resilience annotations.
@Configuration
@EnableResilientMethods
class ResilienceConfiguration
@RetryableUse native @Retryable for straightforward retry behavior.
@Service
class InventoryClient(
private val webClientBuilder: WebClient.Builder,
) {
private val webClient = webClientBuilder
.baseUrl("https://inventory-service.example")
.build()
@Retryable(
includes = [InventoryServiceException::class, WebClientResponseException.ServiceUnavailable::class],
maxRetries = 4,
delay = 200,
multiplier = 2.0,
maxDelay = 2000,
)
fun checkStock(productId: UUID, quantity: Int): Mono<StockResponse> =
webClient.get()
.uri("/api/stock/{productId}?qty={quantity}", productId, quantity)
.retrieve()
.onStatus({ it.is5xxServerError }) {
Mono.error(InventoryServiceException("Inventory temporarily unavailable"))
}
.bodyToMono(StockResponse::class.java)
}
If your public API is suspend, keep the remote client resilience at the reactive boundary and
adapt
at the edge when needed.
suspend fun checkStockAwait(productId: UUID, quantity: Int): StockResponse =
checkStock(productId, quantity).awaitSingle()
@ConcurrencyLimitUse @ConcurrencyLimit to cap concurrent executions of a method when downstream pressure or shared
resource exhaustion is the main risk.
@Service
class NotificationService {
@ConcurrencyLimit(10)
fun sendNotification(request: NotificationRequest): Mono<Void> =
Mono.fromRunnable {
// downstream side effect
}
}
Native Spring resilience is strong for retry and concurrency limiting, but Resilience4j still matters when you need:
@CircuitBreaker@RateLimiter@Bulkhead@TimeLimiter@Service
class PaymentClient(
private val webClientBuilder: WebClient.Builder,
) {
private val webClient = webClientBuilder
.baseUrl("https://payment-gateway.example")
.build()
@CircuitBreaker(name = "paymentGateway", fallbackMethod = "paymentFallback")
fun charge(request: PaymentRequest): Mono<PaymentResponse> =
webClient.post()
.uri("/api/charge")
.bodyValue(request)
.retrieve()
.bodyToMono(PaymentResponse::class.java)
fun paymentFallback(request: PaymentRequest, ex: Throwable): Mono<PaymentResponse> =
Mono.just(
PaymentResponse(
orderId = request.orderId,
status = "PENDING_RETRY",
message = "Payment temporarily unavailable",
),
)
}
@RestController
@RequestMapping("/api/products")
class ProductController(
private val productQueryService: ProductQueryService,
) {
@GetMapping("/search")
@RateLimiter(name = "productSearch")
suspend fun search(@RequestParam query: String): ProductSearchResponse =
productQueryService.search(query)
}
Use layered timeout strategy instead of one giant hammer.
connect-timeout, read-timeout)WebClient config where neededspring:
http:
clients:
connect-timeout: 1s
read-timeout: 2s
Fallbacks must preserve correctness, not just silence errors.
Resilience without observability is theater.
WebClientWebClient| Need | Preferred tool |
|---|---|
| Simple retry | native @Retryable |
| Simple concurrency cap | native @ConcurrencyLimit |
| Stateful circuit breaker | Resilience4j @CircuitBreaker |
| Request rate limiting | Resilience4j @RateLimiter |
| Advanced isolation model | Resilience4j @Bulkhead |
| Declarative timeout layer | Resilience4j @TimeLimiter |
RestTemplate in a reactive stackCompletableFuture as the default resilience model in WebFlux../SKILL.md — Core reactive WebFlux and infrastructure baselinespring-boot-testing-integrations — WireMock, downstream-failure, and adapter integration testsspring-boot-actuator — Metrics, health, and production observability